Hut 8 and IREN Line Up Over $12B in AI Commitments as Public Miners Pivot Beyond Bitcoin

The economics for publicly listed crypto miners are shifting fast, and two new disclosures highlight how quickly AI is moving from side project to core strategy. Hut 8 announced a second 15-year lease tied to its Beacon Point campus in Texas, valued at $9.8 billion. The arrangement effectively dedicates the entire 1 GW site to AI computing for a single tenant, meaning the campus will not be used for Bitcoin mining. In parallel, IREN reported $2.8 billion in new multi-year AI cloud contracts and raised its year-end 2026 annualized AI cloud revenue target to above $4 billion. Taken together, the two updates amount to more than $12 billion in future commitments that sit outside the traditional business of securing the Bitcoin network. A flagship mining build becomes an AI campus Beacon Point was initially positioned as a marquee Bitcoin mining facility. Hut 8's decision to fully allocate the project to AI under a 15-year lease marks a clear strategic reset. The long-term power arrangements and heavy electrical infrastructure that once made the site attractive for ASIC deployments are now being repositioned for GPU-based and specialized AI compute. While the tenant was not named, the structure implies Hut 8 will manage construction and operational scaling, in exchange for a long-duration revenue profile that typical mining operations rarely achieve. A $9.8 billion lease spread over 15 years also helps dampen the balance-sheet volatility miners have faced through multiple halvings and hash price drawdowns. IREN takes a different route: rather than leasing physical capacity, it is selling AI cloud compute. The scale of the $2.8 billion contract haul and the jump in its 2026 target suggest demand is running ahead of expectations. With recurring revenue and high utilization, AI cloud services can offer margins that outpace the economics miners have lived with since early 2024. Implications for hash rate and mining concentration Every megawatt redirected to AI is a megawatt not contributing to Bitcoin's hash rate. The tradeoff is familiar, but the size is not. A gigawatt-class facility like Beacon Point exiting the mining equation for at least 15 years is a meaningful data point. Near-term network security impact appears limited, with global hash rate still near all-time highs. The longer-term signal is a gradual reshaping of the infrastructure base that has supported U.S. mining dominance. As compute-intensive workloads expand across adjacent digital sectors, power, cooling, and interconnection capacity are increasingly contested. That raises the possibility that hash rate becomes more concentrated among pure-play miners that lack the capital or customer access to pivot toward AI. Over the next two years, that could alter both the geographic footprint and corporate mix of Bitcoin mining. Markets have already started to re-rate the story. Hut 8 and IREN have seen their share performance respond more to AI pipeline headlines than to Bitcoin spot moves in recent quarters, as investors weigh them less as single-asset proxies and more as digital infrastructure platforms. Not just Hut 8 and IREN This shift is broader across the public mining cohort. Core Scientific, TeraWulf, and others have also outlined expansion into AI and high-performance computing. The overlap between mining data centers and AI infrastructure is real, but not seamless. AI inference and training typically require higher-grade connectivity, tighter power reliability, and different cooling approaches. Operators able to fund and execute those upgrades are effectively entering a new business with a different investor narrative. The appeal is straightforward: AI cloud revenue is not tied to crypto price cycles, does not face halving dynamics, and is not exposed to network difficulty adjustments. For miners long forced to justify Bitcoin-driven earnings volatility to institutional shareholders, contracted infrastructure income can be an easier story to underwrite. Execution risk remains The pivot comes with operational risk. Building and running large-scale AI data centers demands capabilities that do not fully overlap with managing ASIC fleets. Specialized talent, procurement pipelines, and thermal management requirements are not interchangeable. Long-term AI demand is also a key variable. These agreements represent commitments, but the sector's growth path depends on whether enterprise adoption sustains today's compute buildout. If workloads shift toward more efficient on-device processing, or if regulation slows data center expansion in states such as Texas, the economics could look different within three to five years. Regulatory classification is another open question. As miners morph into hybrid AI infrastructure providers, the compliance perimeter may change. Power purchase agreements, grid interconnection rules, and data center incentives are already being reviewed in several U.S. states. A site that transitions from mining to hosting AI servers for a single enterprise tenant may encounter new local requirements. For now, investors are rewarding the move. Whether these deals deliver the expected revenue over a 15-year horizon will hinge on factors the market cannot fully price today.